Multi-armed bandits for online optimization of language model pre-training: the use case of dynamic masking.
Iñigo UrteagaMoulay-Zaïdane DraïdiaTomer LancewickiShahram KhadiviPublished in: CoRR (2022)
Keyphrases
- language model
- language modeling
- multi armed bandits
- n gram
- online learning
- probabilistic model
- document retrieval
- context sensitive
- speech recognition
- retrieval model
- test collection
- query expansion
- information retrieval
- statistical language models
- language modelling
- ad hoc information retrieval
- language model for information retrieval
- query terms
- relevance model
- document length
- translation model
- training set
- computational complexity
- bayesian networks